KITTI Odometry with OpenCV Python - Pt.1 - Fundamentals (Autonomous Vehicles)

Part 1 of a tutorial series on using the KITTI Odometry dataset with OpenCV and Python. In this video, I review the fundamentals of camera projection matrices, which relate points in 3D space to pixel coordinates, and the role of transformation matrices in relating points in separate coordinate frames. This knowledge lays the foundation for understanding the tools we will use to perform visual odometry using stereo depth estimation and structure from motion using the Perspective-n-Point algorithm with RANSAC.

GitHub repository with Jupyter notebook version of the tutorial:
https://github.com/FoamoftheSea/KITTI...

In Part 2, I code a useful dataset handling class which can be used to easily use and explore the dataset in Python, even on low-RAM computers.

References:
KITTI Odometry Dataset:
http://www.cvlibs.net/datasets/kitti/...
https://www.cs.toronto.edu/~urtasun/p...

Feature Matching in OpenCV:
https://docs.opencv.org/4.0.0/dc/dc3/...
https://docs.opencv.org/4.0.0/da/df5/...

Other great resources on computer vision:
http://www.maths.lth.se/matematiklth/...
https://cmsc426.github.io/sfm/#estfun...
https://www.cc.gatech.edu/classes/AY2...